on Adam Osborne. “It was six months,” he said, “before we ever saw
Osborne’s literature. He introduced his computer in March 1981, but I
didn’t know anything about it.” Some would question that statement, of
course, given the novelty of the Osborne 1 and its creator’s flair for
publicity.
Whatever happened—with Osborne in mind or not—Kay selected the same
basic technology and essentially similar parts.
A Z80-Style Microprocessor
A =microprocessor= is simply the main brains that tie together the
memory devices, the keyboard, and other parts of the computer. You can
also call it a =central processing unit= (=CPU=).
Okay, but why a Z80 in the Kaypro?
That’s simply a style of chip from a company called Zilog. “You could
buy those all over the place,” Kay said. “We wanted something proven and
reliable.” The Z80 was the only kind of CPU that could run =CP/M=
(=Control Program for Microcomputers=)-style programs, then an industry
standard.
The Z80, though, can’t handle programs as complicated as the newer 8088
chip that ended up in IBM PCs and countless clones. Here’s why.
A =bit= is a “1” or a “0.” Patterns of bits form =bytes=. And what’s a
byte? Nothing more than a letter or number. So how does this tie in with
the early Kaypro’s Z80 versus the IBM’s 8088?
Well, the Z80 processes information in =words= of only 8 bits at a time.
The 8088 in the IBM, however, gulps it down in 16-bit chunks and works
with larger memories—so that the IBM can run more complicated programs,
like the more elaborate electronic spreadsheets for forecasting profit
and loss. There’s another benefit, too. A 16-bitter doesn’t need to call
up programming information as often from the spinning floppy disks,
which are slower than the purely electronic memories.
The Kaypro II, however, won’t _immediately_ become a has-bit—er,
has-been. “Converting the Kaypro to a 16-bit machine is no big deal,”
Kay observed. In fact, several months after he said it, his company
announced a computer called the Kaypro II Plus 88 with both the 8-bit
Z80 and the 16-bit 8088. It couldn’t run all IBM programs, especially
those boasting =graphics= or pictures as well as words and numbers.
Here, as in many cases, the term “IBM compatible” is grossly misleading.
There are degrees of compatibility, and even if a computer supposedly
uses the same style of floppy disks, you must test it with the program
_you_ want to run. (Many computer aficionados test for general IBM
compatibility by running a game called Flight Simulator, or Lotus 1-2-3,
a program combining a spreadsheet with graphics and several other
capabilities.)
Some existing programs, by the way, are configured so that it doesn’t
matter much whether you’re running the 8- or 16-bit version.
That will change as special programs come out for the 16-bit machines
and as old ones, like WordStar, begin to take full advantage of 16-bit
capabilities. Already Apple and other companies are writing new software
for computers like the Macintosh, which, if not universally regarded as
a true 32-bit machine, contains chips faster than the IBM’s 8088.
For many small businesses, however, 8 bits may be powerful enough even
for number crunching as long as the right programs are available.
The wisdom in the micro business is that no machine is ever
obsolete—just so it does the job for you.
64K Random-Access Memory (RAM)
RAM, once the power goes off, will throw out everything you’ve typed.
This temporary computer memory, this electronic scratchpad, is like a
bright student cramming for a quiz in a useless subject: he’s a quick
learner—and a quick forgetter. Don’t confuse RAM with =ROM=, or
=read-only memory=. ROMs keep remembering even after the power goes off.
But you can’t “write” new material on them; commonly, ROMs store
programs used over and over again.
What about the RAM, though? How to get around the forgetting? You just
transfer the information to floppy disks, which are slower but as stable
as a music cassette.
Different computers hold different amounts of information in their RAMs.
The Kaypro II holds 64K; =K= is normally a scientific abbreviation for
1,000. So 64K stands for a memory capacity of about 64,000 letters or
numbers. No, I won’t mess with the technical explanation of why it isn’t
exactly 64,000.
Okay, but why _64_K?
It’s a magic number to the computer. Sixty-four K bytes is the largest
memory with which the Z80 and other 8-bit microprocessors can easily
communicate without too much technical trickery.
Also, you need at least 64K to make good use of some popular software
like the WordStar word-processing program.
Moreover, the RAM must be roomy enough to handle electronic
spreadsheets. Most small businesses will find that the 64K Kaypro can
juggle around enough columns and rows on the sheets. But in some cases
you _may_ still want a micro with 500K or even 1,000K or more of RAM. To
sum things up, that’s because:
1. Bigger RAMs can work with more and larger numbers—a handy capability
for a company with extremely complicated spreadsheets.
2. More RAM can accommodate programs more complicated for the computer.
Note the words “for the computer.” Some of the programs most
complicated for the computer may be the easiest for you to use.
Or they may be the most versatile. Lotus 1-2-3, a combination of a
spreadsheet, a graphics program, and several others, is far too
complex for an 8-bit machine with 64K RAM. Many 16-bit programs
require 192K or more just to load the program before you even put in
your reports or spreadsheets.
3. You may want the most sophisticated software to thwart computer
crooks. That could be particularly true in a large company where many
people can poke through a computer’s memory either in person or via
other machines connected by phone or other wires.
Never buy a computer without seeing for yourself that it _and_ the
software will together serve your needs. Don’t take the sales rep’s
word. Instead, make him or his software specialist put the machine
through its paces. Do so even if you’re not buying software at the
store. Of course, the sales rep may not have access to the software, in
which case you may want to visit another store for a demo. And if no one
can assure you, no ifs or buts, that the computer-software combo works
fine? Then you may want to hire a consultant.
Use of CP/M Software
=CP/M= is a kind of =operating system=. It tells a computer how to do
garden-variety jobs, like moving data from a memory unit to one that
does the machine’s electronic reasoning. CP/M isn’t the only operating
system that business microcomputers use. When I was writing this
chapter, however, it was one of the more common, with variation present
on more than half a million computers. A company called Digital Research
marketed CP/M, selling to manufacturers, which, in turn, didn’t have to
create their own operating systems. In the early 1980s, the big software
hits almost always came in CP/M formats.
“We wanted to use CP/M because 10,000 programs were available,” Kay
said. They were for computers with 8 bits, not 16—another reason he made
the original Kaypro an 8-bitter.
“If we’d used 16 bits,” Kay said with uncharacteristic profanity,
“people would have said, ‘Where the hell’s the programs?’ IBM could sell
their computers without programs, which is what they did essentially and
let the programs come later.”
In contracting software design to outsiders, “Big Blue,” as IBM is
called, was defying its own tradition. It hired Microsoft—a Washington
State software company headed by a brilliant young Harvard dropout—to
create a whole new operating system. Called MS-DOS, the system would
become an industry standard by 1984. And from the start, independent
software houses hustled to market MS-DOS-style programs usable on the
hot new machine for applications like word processing and spreadsheets.
The existing CP/M =applications programs= wouldn’t work on an unmodified
IBM Personal Computer unless the programmers converted them to MS-DOS or
another compatible operating system.
‘Two-Disk
They’re the gizmos into which you insert the floppy disks through slots
in the computer.
From the outside a disk looks square. It’s inside a container that helps
protect it from fingerprints and other threats to the plastic disk’s
magnetic coating.
You can’t cover the delicate disk completely, though. Some of it must
touch the =head= of a disk drive. A head “writes” magnetic patterns on
the oxide coating of a disk; in other words, it stores information that
you’ve typed into your computer or the results of your calculations.
And to fetch the information? Then you simply type out commands
instructing the head to “read” from the disk to the RAM.
It’s much like a tape recorder storing your voice magnetically. But a
disk drive can read or write information much more rapidly than a
tape-recorder-style arrangement could. That’s because the disk spins so
rapidly—hundreds of revolutions per minute.
Why twin disk drives, however?
There are two reasons:
1. You can quickly make safety copies of valuable disks—something that’s
more of a hassle with single-drive computers like the bare-boned
Macintosh.
2. You can more easily work with long electronic documents.
With one drive I’m storing WordStar and a number of other goodies that
take up almost all the space on the floppy; there’d be hardly any room
for the information I’m typing out. All this is happening on my “A”
drive.
The “B” drive is my =data disk=, devoted entirely to storage of my
writing. I can easily squeeze the equivalent of about fifty pages of
double-spaced typing with WordStar, which is one reason I bought a
Kaypro instead of an Osborne. Brand O’s earlier versions could store
only half that amount of material. Before buying a computer, you should
always analyze your paper records in terms of “K.” Each K stands for
1,024 characters; and that includes letters, numbers, punctuation marks,
and spaces in between. A double-spaced page of typing might be 66 (the
number of columns across each line) × 28 (the number of lines), or 1,848
characters. Bingo! That’s 1.8K, plus a little slack for safety. Since
Andy Kay allowed 191K of working space on each floppy, I’d have enough
room for a document some hundred pages long, except that WordStar makes
an electronic carbon copy for additional security. So my actual working
space is around 50 pages. That’s easily enough for this chapter to fit
on one of the Kaypro II’s floppies. And newer floppy disk drives can
accommodate more than 2.5 million bytes.
Although the Osborne and Kaypro both used twin drives, there were also
major differences between the two—beyond the =double-density= feature
that allowed Kay to squeeze twice as much material on each floppy.
Osborne’s first computer, for example, came with only a five-inch
monitor. “Adequacy is sufficient,” he once said. “All else is
irrelevant.” It was a neat excuse for his early machine’s limited disk
space and for the five-inch, black-and-white screen, which could display
only 52 columns and which wasn’t as kind to the eyes as green.
The production-model Kaypro, on the other hand, boasted a nine-inch,
80-column green screen from the start.
And the Kaypro keyboard made the Osborne board seem flimsy. “I like
something I can feel when I’m pressing a letter down,” said Kim
DeFilippis, a dealer relations woman who was one of the touch-typists
chosen to try out different keyboards for the new machine. “The board
doesn’t feel like one on a computer. It feels like a typewriter.” And
the keys, unlike the Osborne’s, were in all the right places for people
weaned on Selectrics. Kay had used the same Tiffany-quality keyboard
found on computer terminals costing thousands of dollars more than his
economy machine.
Kay, moreover, had given his Kaypro a metal case unlike the plastic one
of the Osborne. Metal dispersed heat better, and that meant that the
parts might last longer than those in the early Osbornes. Brand O tended
to overheat floppy disks. Recalling that Non-Linear Systems made test
instruments, however, one reviewer gibed that the metal case betrayed
the Kaypro’s heritage. I didn’t care. The Kaypro reminded me of old
Heathkits, of smoky soldering irons, of crisp, cold evenings in my youth
when I’d listened through the static for the whistling dots and dashes
from my fellow hams. But many people shrugged off the Kaypro’s
appearance as unprofessional, and so I asked about the case. Kay said
its contents dictated the shape. “Unless we created air space and made
it larger,” he said, “we couldn’t have made the shape much different.” A
hood shaded the Kaypro’s =cathode ray tube= (CRT; the TV-like screen).
And as it turned out, that matched the comfortable slope of the keyboard
that attached to the front of the machine for carrying. I noticed only a
few adornments added for buyers’ eyes. One was the blue “KAYPRO II”
lettering on the case and the stripes on it and the keyboard; another
was the blue color of the keys on the calculatorlike numbers pad. No,
you couldn’t mistake the Kaypro for a Gucci creation. And yet, as good,
functional design, the Kaypro succeeded: at least one other portable
appeared in a sharp-edged metal box.
The location of my Kaypro’s two disk drives, to the right of the CRT,
was especially logical. Kay and his engineers wanted them well separated
from a high-voltage transformer on the left—an electrical component that
was part of the video circuitry and whose magnetic radiation could
interfere with the operation of the drives. Bill McDonald noted the
Osborne disks were on either side of the tube. Circuitry for a nine-inch
tube produced more electronic noise than that given off by a
five-incher. And McDonald speculated that this was one reason Osborne
had to make do with the smaller size.
Inside the Kaypro, atop the printed circuit boards, you saw integrated
circuits with leglike leads of the kind that one writer compared to
caterpillars. You found somewhat similar looking parts inside the
Osborne. But there was a difference. “Oh, my God,” said Michael Pond,
owner of a Washington-area Kaypro dealership, the Computer Shoppe, “once
you take the cover off the Osborne, everything sort of falls apart in
your hand. It’s held together by the wiring harness. It’s like a
jellyfish.” That may have been a bit off target. Chris Christiansen,
however, a computer analyst with the Yankee Group, a high-tech marketing
research firm, compared the Osborne to a Chinese puzzle and said an
Osborne user “in the next office gets very nervous when I come around
with a screwdriver.” By contrast the Kaypro looked eminently repairable.
“When you take one part of the machine off,” Pond said, “the rest of the
machine doesn’t fall apart.” Andy Kay’s people had divided up the
circuitry into several modules that repairmen could remove _easily_. His
computer had far fewer chips than the old Apple II. And with fewer
chips, repair people more often could locate glitches by replacing
various parts until the machine was working again. Fondly, Pond said the
Kaypro II was “over-engineered,” like an early-model Volvo.
“Well,” you’re wondering, “what about the Kaypro I?”
It never existed except as a prototype. Of course, the II didn’t hurt
Kay’s efforts to convince buyers that he had one up on Osborne.
The prototype’s screen was between the disk drives, interfering with
them, and another flaw became clear: the shortness of the cable between
the keyboard and the main unit. Bill McDonald was worried. Suppose a
customer crushed a finger while wrestling with the little cable? Why not
hinge the keyboard to the twenty-pound console? And that’s how
Non-Linear Systems indeed built the prototype. But McDonald still
fretted that someone might undo the latches; and so, driven by fear of a
lawsuit from a klutzy Kaypro owner, the engineers redesigned the
computer to accommodate a coiled cord several feet long.
Altogether, Andy Kay’s people spent only a year and less than half a
million dollars developing the Kaypro II.
“After we’d been at it eight or nine months,” he said, “someone brought
in some literature from Osborne, who had just announced his portable
computer. I said, ‘Oh, oh, somebody beat me to the punch.’ But I saw he
had a different idea in mind—a smaller screen, plastic case, different
market.”[10] The words “different market” are more than a little off
mark. Andy Kay, like it or not, was in direct competition with Osborne.
Kay might sell his computers at independent stores rather than at the
ComputerLands and Sears stores where many of Osborne’s machines ended
up. But at meetings of Kaypro owners and “Ozzies,” I found the same
hodgepodge of small-business people and professionals.
Footnote 10:
_San Diego_ is the source of the “After we’d been at it eight or nine
months” quote.
Hearing of Osborne, Kay actually was grateful in a sense. “When he
started telling the world about how many orders he’s got, I said, ‘Well,
if he’s getting all these orders, I’ll start making more of them.’”
Kay used an electronic spreadsheet on his newly developed computer to
forecast its sales. The Kaypro II had better succeed, since the half
million in expenses didn’t include marketing costs and miscellaneous
ones such as manuals—another million altogether. Kay, however, said the
high-interest loans for the project “never really bothered me that much.
I felt confident of getting it done. In one sense I place very little
value on money. My wife, Mary, is quite different. She said, ‘You’ll
lose your money.’ She was secretary of the company. She had to sign
mortgages on our home that we had paid for, property we owned, the plant
site. It was a heavy burden for her. If you’re in an airplane or car,
going up a winding road on a mountain, the fellow who’s driving isn’t as
nervous as if someone else is driving. I was involved in getting it
done. She was on the sidelines worrying about it.”
The Kaycomp—that was the original name before the Kays changed it
because it resembled another computer firm’s—first went public at a San
Francisco computer fair in March 1982. Just a routine item appeared in
_Byte_, the phone-book-sized microcomputer magazine. I wondered why.
Maybe it was because Andy Kay’s technology wasn’t new, merely good
repackaging, and he wasn’t selling himself as God or Henry Ford. But
dealers at computer shows were raving. The Kaypro’s suggested retail
price was $1,795, the same as the Osborne’s, and Kay, too, threw in
software: a word processor, an electronic spreadsheet, and other
programs that could have cost more than $1,000 if purchased
individually. Kay was smart. Osborne had pioneered by including business
software for “free” with an economy-priced machine, and now Kay must
follow with its own “bundled” programs.
“Has _Catch 22_’s Milo Minderbinder, World War II’s greatest
wheeler-dealer, hired on as a software buyer at NLS?” _San Diego_
marveled.
“You don’t think that took a lot of time to put together?” Kay said. “We
purchased some software outright and pay royalties for some.”[11]
Footnote 11:
_San Diego_ contains the quote on software purchases and royalties.
By mid-1982, customers and dealers had placed several thousand advanced
orders. Kay’s production lines cranked up, though the pace was slow at
first as his people searched for bugs. They did not always stamp them
out. A disk drive on my Serial #3083, lasted only a year; my warranty
was for the industry’s usual ninety days. Moreover, despite visits to
several dealers, my computer still streaked lightly across the screen
when I typed, and finally I had to have the monitor replaced. But the
view easily beat the Osborne’s. And later, modifying the circuitry and
positioning the disk drives horizontally instead of vertically, Kay
ended the streaks on new units. “In an emergency I get all the senior
engineers on the job,” he said. “We don’t let problems go on like old
man river.”
Andy Kay rewarded his top problem solvers with benefits like stock
options, and the stock offering prospectus from Prudential-Bache
Securities anticipated that in 1983 Kay himself would earn $187,000 in
salary, bonuses, and other remuneration.
Not everyone fared so well; Kay said his labor costs were half those of
competitors. “The wages on the line are so low,” quipped a disgruntled
ex-employee, “I’d call them south of the border.” Kaypro was typical of
many high-tech companies; the production workers were mainly women, many
of them foreign born, some of them incapable of speaking English, all of
them nonunion. Adam Osborne, too, tried to cut labor costs to the bone.
And Atari had laid off scores of Americans and farmed out jobs to
cheaper labor abroad, reddening the faces of the politicians known as
“Atari Democrats” who believed that high tech could fight unemployment.
Regardless of the low wages, Kay’s own company at least appeared to be
the antithesis of a sweatshop. His hillside buildings didn’t look like
normal factories; they were long and narrow, well windowed, split into
small rooms without the racket of mechanized production lines. Kay
described his workers as “always moving, interacting constantly. If one
piece is missing, they work around that. If one person is slow because
he happens to be new, they work around him and help him out. It’s
exactly the same approach we used for stuffing printed circuit boards on
the voltmeters. It uses the least amount of capital equipment, and it’s
the easiest on the assemblers, because they aren’t just sitting or
standing in one spot.”
Lacking a conventional assembly line, Kay said he needed few mid-level
managers; and even after Kay went public, he still hated to bring in
MBAs. Managers built empires. They feuded. They got in producers’ way.
That was how Kay felt, apparently—a legacy of the 1960s when Non-Linear
Systems had splurged thousands on those seven vice-presidents and their
white Cadillacs. But some practices from the go-go years lingered. One,
said Kay, was participative management, the philosophy that had led to
the formation of those small, friendly assembly teams. The atmosphere
around the plant was informal. “We have very few written policies on
anything,” he said. No dress code existed, save for an informal ban on
attire like short shorts—a policy bent to accommodate workers who
labored in the hot Southern California sun.
Also, Andy Kay’s company at times hired people with unusual backgrounds.
Clifford Odendhal, a musician-dancer-songwriter in his thirties, had
come by way of COJO Wind, David Kay’s alternative energy firm. David
asked Odendhal to help Non-Linear Systems catch up on back
correspondence, leading to a customer-service job there. Eventually,
Odendhal was helping with public relations. “This is the first job
outside of entertainment that I have ever cared about beyond my
paycheck,” he said.
Andy said of David, “He’s less cautious about hiring people than I am.
If they don’t work out, he just lets them go.”
David, a six-foot-three-inch surfer with a degree in math and a hobby of
collecting dictionaries, had himself started with the company several
years before the birth of the Kaypro. Now he was vice-president for
marketing, and when an investment banker asked Andy if David was a
vice-president because he was Andy’s son or because he was the best man
for the job, Andy replied: “Both.” He believed that David’s role in the
development and marketing of the Kaypro “was extremely important. In
thirty years I’ve never seen a man in marketing who measures up to
one-tenth the dedication and performance he’s shown.”
Andy’s son used what the _Wall Street Journal_ described as
“idiosyncratic” marketing techniques. Dealers couldn’t sell Kaypros by
mail. They didn’t enjoy extended credit. The Kaypro Division, in fact,
did not even spend much at first on national advertising.